A noninferiority trial comparing the sensitivity and specificity of Tritrichomonas foetus RT-rtPCR using 0.9% sterile saline versus PBS as transport medium

Abstract

The reported limit of detection (LOD) for Tritrichomonas foetus reverse-transcription real-time PCR (RT-rtPCR) is one organism/extraction in PBS without prior culture. Our objective was to determine if 0.9% sterile saline (saline) was noninferior to PBS as a transport medium for T. foetus RT-rtPCR testing at the LOD. Each week for 10 wk, preputial washing samples were collected from known T. foetus–negative bulls, centrifuged, and combined into a single pellet that was evenly distributed into stock solutions for both media. The stock solutions were used to prepare sixty 1.3-mL samples of each medium, with 30 samples inoculated with T. foetus to a concentration of 1 organism/100 μL to evaluate sensitivity (Sn) and 30 left uninoculated to evaluate specificity (Sp). We tested 1,200 (Sn_PBS = 300, Sn_saline = 300, Sp_PBS = 300, Sp_saline = 300) samples by RT-rtPCR. Differences in Sn and Sp between media were tested using logistic regression models. The Farrington–Manning test was used to evaluate the difference between parameters with ∆ = 2.5%. For PBS, Sn = 70.7% (95% CI [65.2, 75.8%]) and Sp = 99.7% (95% CI [98.2, 100%]); for saline, Sn = 73.3% (95% CI [68.0, 78.3%]) and Sp = 100% (95% CI [98.8, 100%]). The Sp_saline was noninferior to Sp_PBS (p = 0.002); however, noninferiority of Sn_saline to Sn_PBS was inconclusive (p = 0.08). The Sn for samples inoculated near the LOD in saline was similar to, and Sp was noninferior to, PBS, suggesting that either would be acceptable as a transport medium for T. foetus detection using RT-rtPCR.